英文:
What is the exact mathematical formulation of calculation of Mutual Information and F-Score in scikit-learn?
问题
我想了解两个指标 Mutual Information (MI) 和 F-score 在这个 scikit-learn 示例 中的确切数学表达式。文档仅陈述如下：
- 对于 MI：参考了几篇论文和维基百科页面，但没有细节。
- 对于 F-score：
- 使用 r_regression 计算每个回归器与目标之间的交叉相关性，如下图所示。
- 将其转换为 F 分数，然后转换为 p 值。
英文:
I would like to know the exact mathematical formulation of the two metrics Mutual Information (MI) and F-score used in this example in scikit-learn. The documentation states only the following:
- for MI: referring to couple of papers and the wikipedia page without details.
- for F-score:
- The cross correlation between each regressor and the target is computed using r_regression as:
- It is converted to an F score and then to a p-value.
答案1
得分: 1
对于scikit-learn中的互信息（Mutual Information，MI）和F分数（F-score）的计算，可能会根据变量是分类还是数值型而有所不同。
您可能会发现直接查看源代码有所帮助：
- 对于互信息（Mutual Information）：
- 用于分类任务的：
mutual_info_classif - 用于回归任务的：
mutual_info_regression
- 用于分类任务的：
- 对于F分数（F-score）：
f_regression
希望这能帮助您。
英文:
The calculations for Mutual Information (MI) and the F-score in scikit-learn may vary based on whether the variables are categorical or numeric.
You might find it helpful to look directly at the source code:
- For Mutual Information:
- For classification tasks:
mutual_info_classif - For regression tasks:
mutual_info_regression
- For classification tasks:
- For the F-score:
f_regression
I hope this helps.
通过集体智慧和协作来改善编程学习和解决问题的方式。致力于成为全球开发者共同参与的知识库，让每个人都能够通过互相帮助和分享经验来进步。
评论